-
We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.
OFFICIAL: Post your 3dMark 2003 scores here.
Page 7 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Heh increased the core to 350 and the memory to 325, and it upped my score 360points to 5087 . However a 300mhz CPU increase/decrease yields a point difference of ~100?? 😕 This benchmark is BROKEN. I'd probably get another 300 by increasing it to 375/350, its that predictable. I'm willing to bet I could toss my 9700pro into my P3 500 rig (if my trusty ole BE6-II could handle it 🙂 ) and score higher than some of the non-DX9 part scores with a CPU that's 5-10x as fast in real life apps.
This "benchmark" heavily weighs tests that a) aren't existent in today's (or any in the near future) games and b) probably won't ever be. Plus its plain ugly 😛 The best looking test was the DX7 bench for chrissake. Its supposed to be "The Ultimate Gamer's Benchmark", but it completely ignores the fact that game performance relies on all north bridge subsystems, not just the GPU. And to think many of us bashed 3dmark2k1 for being a worthless benchmark 😉
Chiz
Edit: Btw, I could easily see it was GPU dependent based on my CPU testing scores in comparison to those that scored higher than me. My CPU marks were much higher, yet they had 300-600 higher point scores. The ONLY differences were driver (3.1) and vid card clockspeeds (I ran at defaults for all settings). What other tweaks that were made, don't know, don't care.
This "benchmark" heavily weighs tests that a) aren't existent in today's (or any in the near future) games and b) probably won't ever be. Plus its plain ugly 😛 The best looking test was the DX7 bench for chrissake. Its supposed to be "The Ultimate Gamer's Benchmark", but it completely ignores the fact that game performance relies on all north bridge subsystems, not just the GPU. And to think many of us bashed 3dmark2k1 for being a worthless benchmark 😉
Chiz
Edit: Btw, I could easily see it was GPU dependent based on my CPU testing scores in comparison to those that scored higher than me. My CPU marks were much higher, yet they had 300-600 higher point scores. The ONLY differences were driver (3.1) and vid card clockspeeds (I ran at defaults for all settings). What other tweaks that were made, don't know, don't care.
Ok first time i Ran it 1916 running at stock speeds (9500 non pro with softmod to 9700)
seemed low compared to other people, so hmm upped the clocks from 275/275 to 350/310
got a mighty 2010 marks. Odd i thought, then realised im an idiot and had left on AA and AF whoops. Still getting used to not using my Geforce 2 MX lol.
Anyway I have to agree with Chiz on this one after much fiddlin with settings and testing with different FSB's, CPU clocks and GPU clocks it appears only the GPU has any kind of effect whatsoever in real terms. A 300mhz difference in my Athlon only fluctuated my score by about 60 points.
However after changing from 275 / 275 clock and mem speeds for a score of 4046
to 357 / 310 speeds I jumped to a huge 4846 marks.
After seeing the performance of the Ti4200's I have to say i find this result a little meaningless. Yes 9500 has DX9 capability but 4x more capable of running games than a decent Ti4200 hmmmm I don't think so some how. Makes me wonder what the FX would get.
if no one has been yet go to HardOCP.com and look for their review of this benchmark. Pretty water, little value seems to be the consensus on this. So anyone with a low score and a decent system..... why worry about it :-D
seemed low compared to other people, so hmm upped the clocks from 275/275 to 350/310
got a mighty 2010 marks. Odd i thought, then realised im an idiot and had left on AA and AF whoops. Still getting used to not using my Geforce 2 MX lol.
Anyway I have to agree with Chiz on this one after much fiddlin with settings and testing with different FSB's, CPU clocks and GPU clocks it appears only the GPU has any kind of effect whatsoever in real terms. A 300mhz difference in my Athlon only fluctuated my score by about 60 points.
However after changing from 275 / 275 clock and mem speeds for a score of 4046
to 357 / 310 speeds I jumped to a huge 4846 marks.
After seeing the performance of the Ti4200's I have to say i find this result a little meaningless. Yes 9500 has DX9 capability but 4x more capable of running games than a decent Ti4200 hmmmm I don't think so some how. Makes me wonder what the FX would get.
if no one has been yet go to HardOCP.com and look for their review of this benchmark. Pretty water, little value seems to be the consensus on this. So anyone with a low score and a decent system..... why worry about it :-D
human2k, I am using Wizzard's hacked 6292/Cat31 ati2mtag.sys and can confirm it doubles the fillrate. You may want to try again or use RivaTuner's patch script on your own file. You should run only the fillrate test to compare original and hacked of the same set.
I still have not been able to submit a score. On FM's message board a user claims certain ranges of addresses are blocked. Mayhaps they plan to progressively increase access to prevent meltdown.
Achtung! Radeon owners should check the status of HyperZ. I found it was disabled. At least it was after installing Rage3D Tweak's tweaker component which I have not used for awhile but I did not check the status before.
Example:
[HKEY_LOCAL_MACHINE\SYSTEM\ControlSet002\Services\ati2mtag\Device0]
"DisableHyperZ"="1"
The value needs to be changed to "0". The location may vary by ControlSet number so just search and change all instances and you should reap a noticeable boost in performance. Having to manually enable it brings back fond memories of my Radeon LE!
My latest score is 4478.
Celeron 1.1 @ 1500
Radeon 9500 @ 9700
Still whoopin' some of y'all even with 9700's but faster CPU's and RAM so get to work tweaking, or untweaking as the case may be! 😉
I still have not been able to submit a score. On FM's message board a user claims certain ranges of addresses are blocked. Mayhaps they plan to progressively increase access to prevent meltdown.
Achtung! Radeon owners should check the status of HyperZ. I found it was disabled. At least it was after installing Rage3D Tweak's tweaker component which I have not used for awhile but I did not check the status before.
Example:
[HKEY_LOCAL_MACHINE\SYSTEM\ControlSet002\Services\ati2mtag\Device0]
"DisableHyperZ"="1"
The value needs to be changed to "0". The location may vary by ControlSet number so just search and change all instances and you should reap a noticeable boost in performance. Having to manually enable it brings back fond memories of my Radeon LE!
My latest score is 4478.
Celeron 1.1 @ 1500
Radeon 9500 @ 9700
Still whoopin' some of y'all even with 9700's but faster CPU's and RAM so get to work tweaking, or untweaking as the case may be! 😉
Yah submission success rates right now are still pretty poor, I can only get em uploaded late EST or early in the morning before I go to work. I've been running a bench or two before bed and then I'll run one overnight and upload in the am.
By the way, total uploaded projects is capped at 5!!! They're really pushing to get people to pay I guess.
I also noticed HyperZ was disabled, but isn't there an option for it in Rage3d Tweak? I'll have to check it out. I leave everything at application preference or defaults for the most part though. I normally don't use any tweakers, but since I can't get 3.1's CP to work, it was the only way I could change AA and AF settings.
Chiz
My biggest problem with this as a benchmark is that it just way too synthetic. They tried to add some cpu to it with the physics calculations in the 2nd (guys falling over when hit...pre scripted?) and 3rd scene (ogre hitting the door, hair) for example, but as shown in the previous threads, increasing the cpu speed does not speed up those calculations and let the gpu do its job. In all current games, the biggest limiting factor of fps is the cpu. Take 1024x768...even in ut2k3 right now a geforce fx, 4 ti series and radeon 9500+'s al get about the same fps until fsaa and AF are added. Sure, its great as a benchmark to see how cards will possibly perform in future games, but nvidia even comments that in game tests 2 and 3 that the rendering method is programmed using inefficient techniques and that it also falls back on 1.1 shaders if 1.4 shaders are not present. While I do not know the specs of 1.3 shaders, they wouldnt be a different version than 1.1 if they did not offer some improvment as 1.4 offers nearly a 40% reduction in polygons rendered per scene over 1.1. EIther way, through my babbling I think it is funny that my score of 1600 on an athlon xp @ 1580 mHz, geforce 4 ti 4200 @ 320/610 and 512 of crappy pc2100 cas 2.5 ram is beating a bunch of systems with processors around 2.5 gHz.
Ketchup
Elite Member
I am sure people said the same thing about 3DMark 2001 when it came out (heck, some people still do say that.) The problem is that this benchmark is going to last a year or two before it is replaced, so it makes sense that it is video card limited at this point, since most video cards don't have the raw bandwidth or technology to run this benchmark correctly. Once it has been around for a while, I think you will see this benchmark scales with the CPU and the video card performance like its predecessors did.
Aaaah... now I don't feel so bad.
Turns out that the new 3DMark was orchestrated to favor DX9 cards (ATI must have donated a lot of money...) 🙂
Then, I'm just fine now even with the low score.
I'd suggest the same for the other low scorers.
In a few months, there'll be opportunity to buy a DX9 card (NV30 or R350??). For now, it's not needed (yet!)
Save the money now and maybe you could buy the next card just from the interest 😉
Turns out that the new 3DMark was orchestrated to favor DX9 cards (ATI must have donated a lot of money...) 🙂
Then, I'm just fine now even with the low score.
I'd suggest the same for the other low scorers.
In a few months, there'll be opportunity to buy a DX9 card (NV30 or R350??). For now, it's not needed (yet!)
Save the money now and maybe you could buy the next card just from the interest 😉
obeseotron
Golden Member
With the Ti4400 tweaked to the max (290/689) I got 1674.
chizow,
Yeah there is an option for HyperZ in Rage3D Tweak but like ye I normally leave graphics setting alone except for forcing AF and Vysnc. I had not actually installed the tweaker bit in awhile and had no reason (or so I thought) to look at the settings in the registry so was surprised to find it disabled.
Update: Ichneumon sez HyperZ is enabled by default and the reading in Rage3D Tweak is a mistake and does not reflect or alter the default but only shows what it should be. I really appreciate the work they have put into R3DT but that is not a good thing. I am going to reinstall the drivers, CP, and R3DT without the tweaker component again (just Custom Display Modes & Overclocker with The Nothing skin).
As for your CP troubles have you tried:
uninstall Rage3D Tweak, CP & driver
reboot
cancel auto install
delete any left over ATI system files (not usually any)
delete ATI, Registry Tweaker & GameUtil reg entries with jv16.org's RegCleaner
install driver & CP
reboot
Yeah there is an option for HyperZ in Rage3D Tweak but like ye I normally leave graphics setting alone except for forcing AF and Vysnc. I had not actually installed the tweaker bit in awhile and had no reason (or so I thought) to look at the settings in the registry so was surprised to find it disabled.
Update: Ichneumon sez HyperZ is enabled by default and the reading in Rage3D Tweak is a mistake and does not reflect or alter the default but only shows what it should be. I really appreciate the work they have put into R3DT but that is not a good thing. I am going to reinstall the drivers, CP, and R3DT without the tweaker component again (just Custom Display Modes & Overclocker with The Nothing skin).
As for your CP troubles have you tried:
uninstall Rage3D Tweak, CP & driver
reboot
cancel auto install
delete any left over ATI system files (not usually any)
delete ATI, Registry Tweaker & GameUtil reg entries with jv16.org's RegCleaner
install driver & CP
reboot
Using the Moobox which contains a Radeon 8500 and 1Ghz@1.2GHz TBird:
16x AF Enabled
VSync On
1024x768x32
827 3DMarks
My score used to be a little above 8000 in 2001SE. I'm glad the heat was turned up, but everyone runs at like 10FPS or less (Except the DX7 game). It's crazy!
16x AF Enabled
VSync On
1024x768x32
827 3DMarks
My score used to be a little above 8000 in 2001SE. I'm glad the heat was turned up, but everyone runs at like 10FPS or less (Except the DX7 game). It's crazy!
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 25K
-
-
-
